Dear Tom,
After much discussion, we've decided not to change the current behavior of the system. We don't believe the potential cost of recognizing "breaks" outweighs the benefits.
But we do recognize there are benefits. Here in NetApp Engineering, we put machines on remote power strips, so if a filer really isn't responding to the console at all we can yank its power. As others have noted, you can interrupt a boot sequence in the firmware by typing <del> during the POST and initialization sequence; you get a nice pliable ok prompt. The combination of terminal servers and remote-controlled power strips gives us enough functionality that we don't notice the lack of <break>. And yes, we're sensitive to the issues of remote administration because that bites us too: I'd much rather develop and test in my office, with my coffee warmer and my headphones, than up in a crowded, chilly, noisy lab.
Regarding net booting: we've had issues with functionality, testing and documentation, and it's never been a high enough priority to steal resources from other projects. But we've started drafting requirements for formally supporting net booting in the future, and I believe we will eventually officially document its use. I can't tell you when because I honestly don't know.
Yours, Mike Tuciarone Platform Software